Hello All, I've a dataset of six samples and 1530 variables/features and wish to know the the importance of features. I'm trying to use the "Rank Features By Importance" as mentioned in Feature Selection with the Caret R Package (http://machinelearningmastery.com/feature-selection-with-the-caret-r-package/) I'm using the following code: rm(list=ls()) set.seed(12345) library(mlbench) library(caret) options(error=utils::recover) #Pastebin link for Data: http://pastebin.com/raw/cg0Kiueq mydata.df <- read.table("data.PasteBin.txt", header=TRUE,sep="\t",stringsAsFactors=TRUE) dim(mydata.df) lvq.control <- trainControl(method="LOOCV") lvq.model <- train(ID~., data=mydata.df, method="lvq", trControl=lvq.control ) #FAILS importance <- varImp(lvq.model, scale=FALSE) print(importance) plot(importance) The data can be downloaded from the following Pastebin link: http://pastebin.com/raw/cg0Kiueq
The program fails to execute with the following error and debug messages: Error in seeds[[num_rs + 1L]] : subscript out of bounds 1: train(ID ~ ., data = mydata.df, method = "lvq", trControl = lvq.control) 2: train.formula(ID ~ ., data = mydata.df, method = "lvq", trControl = lvq.con 3: train(x, y, weights = w, ...) 4: train.default(x, y, weights = w, ...) I've read from multiple sources (http://davidhughjones.blogspot.com/2015/04/r-tip-caret-error.html) that unless the response variable is of class factor Caret issues error like this. However, my response variable('ID') is indeed a factor > str(mydata.df$ID) Factor w/ 2 levels "NONRC","RC": 2 2 1 1 2 1 The detail of my version of R and Caret are as follows: > packageVersion("caret") [1] ‘6.0.70’ R version 3.3.0 (2016-05-03) Platform: x86_64-w64-mingw32/x64 (64-bit) Running under: Windows 7 x64 (build 7601) Service Pack 1 Can someone please suggest any remedy? Thanks in advance [[alternative HTML version deleted]] ______________________________________________ R-help@r-project.org mailing list -- To UNSUBSCRIBE and more, see https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
